Context and question swapped in the code snippet

by J4Q8 - opened May 24, 2023

Discussion

J4Q8

May 24, 2023

•

edited May 24, 2023

Hi,

Thank you for sharing this model! I have been playing with it and noticed that in your "prepare_answering_input" function you have swapped context with question. After putting the context first and question second (which is the way you describe it in text) the probability distribution improved significantly to 0.99 for the correct option.

Just wanted to point it out so that your docs are correct.

Best regards,
Jakub

potsawee

Owner May 27, 2023

Hi @J4Q8 ,

Thank you very much for pointing out the mistake in my example. You are correct that I swapped context & question in "prepare_answering_input".

I've fixed this issue in README.

Best regards,
Potsawee

potsawee changed discussion status to closed May 27, 2023

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment